Utility Proportional Fair Bandwidth Allocation - An Optimization Oriented Approach
Please always quote using this URN: urn:nbn:de:0297-zib-8079
- In this paper, we present a novel approach to the congestion control and resource allocation problem of elastic and real-time traffic in telecommunication networks. With the concept of utility functions, where each source uses a utility function to evaluate the benefit from achieving a transmission rate, we interpret the resource allocation problem as a global optimization problem. The solution to this problem is characterized by a new fairness criterion, \e{utility proportional fairness}. We argue that it is an application level performance measure, i.e. the utility that should be shared fairly among users. As a result of our analysis, we obtain congestion control laws at links and sources that are globally stable and provide a utility proportional fair resource allocation in equilibrium. We show that a utility proportional fair resource allocation also ensures utility max-min fairness for all users sharing a single path in the network. As a special case of our framework, we incorporate utility max-min fairness for the entire network. To implement our approach, neither per-flow state at the routers nor explicit feedback beside ECN (Explicit Congestion Notification) from the routers to the end-systems is required.
Author: | Tobias Harks |
---|---|
Document Type: | ZIB-Report |
Tag: | Utility proportional fairness; congestion control; real-time applications; resource allocation |
MSC-Classification: | 90-XX OPERATIONS RESEARCH, MATHEMATICAL PROGRAMMING / 90Bxx Operations research and management science |
Date of first Publication: | 2004/08/10 |
Series (Serial Number): | ZIB-Report (04-32) |
ZIB-Reportnumber: | 04-32 |
Published in: | Appeared in: Proceedings of QoS in Multiservice IP Networks 2005, Catania, Italy. LNCS vol. 3375, pp 61-74 |